JP2012515538A - Prognostic diagnosis of breast cancer patients by observing the expression of two genes - Google Patents

Abstract

The present invention relates to the expression of two genes, cyclin G2 and Sharp1, which correlate with prognosis in individuals with breast cancer. Specifically, the present invention provides a method of stratifying a sample from a breast cancer patient into high or low risk of recurrence several years after removal of the primary tumor. This classification can be achieved by analysis of protein or mRNA expression levels for two identified genes. The present invention also demonstrates how cyclin G2 and Sharp 1 are identified in breast adenocarcinoma cell lines and validated as effective metastasis predictors in a large cohort of human patients.
【Selection chart】 None

Description

Detailed Description of the Invention

Field of the Invention
The present invention relates to minimal gene signatures that provide information useful for the recurrence of breast cancer by molecular methods based on nucleic acid or protein levels.

[Background technology]
Breast cancer is the most common cancer in women. In the United States, one in eight women is estimated to develop some type of breast cancer by age 85.

  Although the tumorigenic mechanism of most breast carcinomas is largely unknown, there are genetic factors that can predispose some women to develop breast cancer (Miki et al., 1994). More recently, the discovery and characterization of BRCA1 and BRCA2 has broadened our knowledge of genetic factors that can contribute to familial breast cancer, but only about 5-10% of breast cancers are associated with BRCA1 and BRCA2 It is. BRCA1 is a tumor suppressor gene involved in DNA repair and cell cycle control that is important for maintaining genomic stability. Like BRCA1, BRCA2 is also involved in the development of breast cancer and plays a role in DNA repair, but unlike BRCA1, it does not participate in ovarian cancer. Other genes, such as c-erb-2 (HER2) and p53, have also been linked to breast cancer (Beenken et al., 2001). Overexpression of c-erb-2 (HER2) and p53 was correlated with poor prognosis.

  However, to date, none of the other clinically useful markers consistently associated with breast cancer are currently associated with sporadic tumors, ie with known germline mutations that make up the majority of breast cancer The tumor has not been confirmed.

  In clinical practice, accurate diagnosis of various subtypes of breast cancer is important, as treatment options, prognosis and potential for treatment response all vary widely with diagnosis. In breast cancer, early diagnosis and risk stratification are extremely important because breast cancer morbidity and mortality increase significantly if not detected early in the progression of the cancer. An accurate diagnosis or determination of distant metastasis free survival allows oncologists to tailor the application of adjuvant chemotherapy to the most aggressive treatment of women with worse prognosis. It will be possible. Furthermore, since patients likely to be studied can be stratified according to prognosis, accurate prediction of poor prognosis will have a major impact on clinical trials of new breast cancer therapies.

  Typically, diagnosis of breast cancer requires histopathologic evidence of the presence of a tumor. In addition to diagnosis, histopathological examination also provides information on prognosis and choice of treatment plan. Prognosis may also be established based on clinical parameters such as tumor size, tumor grade, patient age, and lymph node colonization by tumor cells.

  The diagnosis and / or prognosis can be determined by direct examination of the outside of the breast or through mammography or other radiographic methods, although to varying degrees of efficacy. However, the latter approach involves considerable social and personal costs.

  More recently, the FDA has microarrays of cDNAs of more than 70 genes measured in fresh or frozen breast cancer biopsies based on the published work of van't Veer (van't Veer et al., 2002) The analysis has approved the gene expression profiling test system for breast cancer prognosis, MammaPrint®.

  This test is for physicians only, but again it needs to be performed with a special measurement tool such as a DNA bioanalyzer / microarray scanner. This is a major drawback, as the results can only be provided by large hospitals or companies that have developed means and standard procedures to perform such complex analyses.

  From the above, the advantages of the present invention based on predictive prognostic value of expression analysis of as few as two genes can be easily understood.

  Simultaneous analysis of 10 genes indeed requires array technology, but simple evaluation of the expression of cyclin (Cyclin) G2 (CCNG2) and Sharp1 (BHLHB3, BHLHE41) does not require array technology. Viewed from another aspect, standard breast cancer prognosis methods, such as primary mass assessment, lymph node metastasis and cancer staging, are currently inadequate to predict disease progression. is there. The disease, especially in the case of cancers defined as medium-aggressive by the canonical criteria, by combining conventional histological methods and the molecular characterization of the tumor with this minimal signature. A sophisticated and inexpensive way to predict progression and recurrence risk will be possible.

［式中、
サイクリンＧ２を単独で用いる場合にはＫ＝１であり、サイクリンＧ２及びＳｈａｒｐ１の両方を用いる場合にはＫ＝２であり、

は未知試料ｉ中のサイクリンＧ２又はＳｈａｒｐ１の発現レベルであり、

はそれぞれ、病歴が知られている母集団におけるサイクリンＧ２及び／又はＳｈａｒｐ１の発現レベルの推定平均値及び推定標準偏差値を表す］
と定義され、０未満のシグネチャースコアは乳癌再発リスクの増大を示す。 [Summary of the Invention]
The present invention comprises the step of detecting cyclin G2 (Gene ID (Gene ID) = 901) gene expression levels in a sample alone or in combination with Sharp 1 (Gene ID = 79365) gene expression levels, the risk of recurrence of breast cancer patients On the evaluation method of This detection includes the steps of measuring a signal directly involved in the expression of the gene (s) in the sample, obtaining the signal, and assessing the cancer recurrence risk of the breast cancer patient. The step of assessing the risk of cancer recurrence in breast cancer patients is performed by calculating the signature score of the cyclin G2 gene expression value alone in the unknown sample or preferably both cyclin G2 and Sharp1 gene expression values Signature score is

[In the formula,
K = 1 if cyclin G2 is used alone, K = 2 if both cyclin G2 and Sharp1 are used,

Is the expression level of cyclin G2 or Sharp1 in unknown sample i,

Respectively represent estimated mean and estimated standard deviation values of expression levels of cyclin G2 and / or Sharp1 in a population with a known medical history]
And a signature score less than 0 indicates an increased risk of breast cancer recurrence.

  The detection can be performed by molecular and / or immunological means. In this case, molecular means mean nucleic acid based assays such as PCR, microarray analysis or Northern blot.

This method is
Quality control of acquired signals;
Normalizing the signal,
Optionally, the method further comprises statistical analysis of the signal by rescaling the acquired signal, preferably performed by computer implemented software.

The present invention is a kit for evaluating cyclin G2 expression alone or in combination with Sharp1 expression in a sample from a breast cancer patient and measuring the risk of breast cancer recurrence, preferably a cyclin G2 specific reagent, preferably SEQ ID NO: An oligonucleotide comprising a 13-mer oligonucleotide derived from 1 or an oligonucleotide comprising at least a complementary sequence thereof,
An oligonucleotide according to a Sharp 1 specific reagent, preferably an oligonucleotide comprising at least a 13-mer oligonucleotide derived from SEQ ID NO: 2 or a complementary sequence thereof
The signature score of the unknown sample is calculated according to the calculations defined for the method, and the unknown sample is in the minimum signature low group if the signature score is negative or the minimum signature height group if the signature score is positive Instructions for use in classification, including classification into the minimum signature low group, further provide a kit indicating a high risk of cancer recurrence in breast cancer patients.

  According to a preferred embodiment, the instructions are implemented by software. Optionally, the kit can further comprise cyclin G2 and Sharp 1 standard expression control high and low as expression levels or nucleic acid samples as a reference standard. Said expression value or nucleic acid sample is preferably derived from a non-metastatic breast cancer cell line and / or a highly metastatic cell line, respectively.

FIG. 6 shows that mutant p53 expression promotes TGFβ pro-migratory response. (A) H1299 cell lysate-Western blot of parental, ie p53 deficient expression (defect) or mutant p53 (p53 R175H). As observed by Smad3 phosphorylation (P-Smad3), the TGFβ signaling cascade is equally active in any cell line. Lamin B (Lamin-B) is a loading control. (B) Effects of TGFβ (TGFβ 5 ng / ml, 24 hours) on the morphology of H1299 cells. (C) Wound healing assay of H1299 cells showing the effect of mutant p53 on TGFβ driven migration. Photographs were taken 30 hours after culture scratching. (D) H1299 cells were seeded on transwell membranes. When indicated as TGFβ, cells were treated with TGFβ (4 ng / ml). The graph shows the number of cells migrated through the transwell after 16 hours. Only H1299 cells reconstituted with p53R175H have acquired the ability to migrate in response to TGFβ. FIG. 5 shows that in breast cancer mda-mb-231 cells, mutant p53 is required for TGFβ-driven invasion and metastasis. (A) Western blot showing p53 protein depletion in MDA-MB-231 (MDA-shp53) expressing shRNA targeting p53. MDA shGFP is a control cell line. (B) Transwell assay of TGFβ-dependent migration of MDA-MB-231 cell line. This response is dependent on canonical Smad signaling, as evidenced by Smad4 deprivation that occurs following blockade of migration. The endogenous mutant p53, which is expressed from these natural loci in these cells, is required for this action. (C) Assay of infiltration activity of MDA-MB-231 cells embedded in 1 drop of Matrigel. The panels show photographs of the same field at different points in time. The dotted lines highlight the edge of the Matrigel droplet. Only control cells have the ability to evade Matrigel® (arrows). This process is blocked by treatment with the TGFβR1 inhibitor SB431542 (5 μM) and thus relies on TGFβ signaling. MDA shp53 cells have impaired function for substrate degradation and avoidance. (D) MDA-MB-231 cells, once embedded in Matrigel®, exhibit a spindle shape in 3D culture conditions (upper panel). The arrowhead shows the process of lamellipodia. Conversely, MDA shp53 formed clusters of adherent cobble-like cells (lower panel). Inhibition of TGFβ signaling is consistent with the phenotypic effects of mutant p53 deficiency (data not shown). (E and F) Fat bodies of SCID mice were injected with MDA shGFP or MDA shp53 cells. (E) Growth rates of primary tumors were similar between the two cell populations. (F) Number of mice with positive lymph node metastasis score. (G, H and I) Lung colony formation assay after injection of the MDA-MB-231 cell line into the tail vein (n = 10 mice of each cell line, 1 × 10 ⁶ cells / mouse). The panels show immunohistochemical staining representative of human cytokeratin in lung sections from mice injected with MDA shGFP (G) or MDA shp53 (H). (I) The graph quantifies lung parenchymal infiltration by control (shGFP) and two independent MDA shp53 clone cell lines. FIG. 5 shows the identification of a novel candidate metastasis suppressor downstream of TGFβ / mutant-p53 in metastatic breast cancer cells. (A) Summary of TGFβ target genes from microarray analysis of MDA-MB-231 cells. The graph shows functional classification of genes controlled by TGFβ in both MDA shGFP and MDA shp53 cell lines. Many genes encode proteins involved in cell invasion, migration and metastasis ("invasion program"). (B) Genes of MDA-MB-231 cells co-regulated by TGFβ and mutant p53. The table shows TGFβ induction levels for the indicated genes from microarray expression data. The difference in fold induction between MDA shGFP and MDA shp53 samples is statistically significant, as indicated by the q value. (C) Verification by Northern blot of ADAMTS9, Sharp1, cyclin G2, follistatin and GPR87 as mutated p53 dependent targets of TGFβ in MDA-MB-231. When indicated as (+), cells were treated with TGFβ1 for 2 hours. GAPDH is a loading control. (D) Regulation of Sharp1 and cyclin G2 expression by TGFβ and mutant p53 in MDA-MB-231 cells. Northern blot analysis of MDA shGFP and MDA shp53 cells treated or not for 2 hours with TGFβ1. GAPDH is a loading control. Both genes are downregulated by TGFβ in control cells but not after mutant p53 knockdown. (E) Sharp1 and cyclin G2 are important effectors of TGFβ / mutated p53 in migration control. Transwell migration assay of MDA-MB-231 cells transiently transformed with the indicated siRNA. The impairment of TGFβ-driven migration in mutant p53 deficient cells can be rescued by the simultaneous deletion of Sharp1 or cyclin G2. β-actin is a loading control. FIG. 1 shows clinical validation of minimal signature as a likely predictor of breast cancer recurrence. Verification of the predictive power of the minimal signature (Sharp1 + Cyclin G2) on a panel of 5 independent data sets summarizing more than 940 tumors (see Table 3 for a full description of these data). The NKI data set (see FIG. 6) was analyzed separately. These analyzes divide tumor samples into two groups by consistent low or high expression of both genes visualized by box plot graphs. "Low" (blue) and "high" (red) are the designations for the minimum signature low group and the minimum signature high group, respectively. The Kaplan-Meier graph on the left indicates that in the breast cancer data set analyzed, patients stratified according to the minimal signature may remain metastasis free, or may remain relapse free or disease free. ing. Logrank test p-values show a significant association between minimum signature height and longer survival. Similar results were obtained using the unsupervised clustering method, which yields the smallest signature low group and the smallest signature high group (data not shown). On the right, Kaplan-Meier survival graphs from the same tumor data stratified according to the 70 gene signature (van't Veer et al., 2002) for comparison. FIG. 6 shows that the minimal signature is associated with both bone and lung distant metastasis risk. Kaplan-Meier curves indicate that MSK samples (Minn et al., 2005) stratified according to the minimal signature may remain without lung (left) and bone (right) metastasis. The minimal signature has statistically significant predictive power for both organ specific metastatic events. Analysis of cyclin G2 expression is sufficient to predict metastasis free survival in the NKI data set. In the NKI data set (sample number 295), as few as one cyclin G2 expression data can be used to classify tumors according to metastatic tendency. Since Sharp 1 expression data is not available from the NKI data set, we set the threshold for cyclin G2 expression based on the proportion of patients with good prognosis (see “Experimental Procedures” for details). Box plots for cyclin G2 and Kaplan-Meier metastasis-free survival curves are obtained using this threshold. FIG. 5 shows that the minimal signature divides grade 2 tumors into two groups with different outcomes. Kaplan-Meier indicates the possibility that the patients from the Stockholm, Uppsala and NKI datasets stratified according to the Nottingham histologic scale will remain free of recurrence, remain free of disease or remain free of metastasis. Curves (grade 1 dotted line, grade 2 violet color line, grade 3 dashed line). Grade 2 tumors (solid line) were further divided into two groups by applying a minimal signature (red line-grade 2 and minimal signature height, blue line-grade 2 and minimal signature low). It should be noted that the high and low groups, respectively, exhibited recurrence free survival rates similar to Grade 1 or Grade 3 patients.

Definitions and Abbreviations Cyclin G2, also referred to as CCNG2, is identified by gene ID = 901 (SEQ ID NO: 1). Sharp1, also referred to as DEC2, BHLHB3, BHLHE41 (containing a basic helix loop helix domain) is identified by gene ID = 79365 (SEQ ID NO: 2).

Template A template of minimal signature is obtained by measuring expression levels of cyclin G2 alone or preferably in combination with Sharp1 in a population of tumor samples from patients with a known medical history.

  Templates are calculated for each of the different assays used to determine cyclin G2 and Sharp1 expression measures.

即ち、母集団又はデータセットにおけるサイクリンＧ２発現レベル及び好ましくはＳｈａｒｐ１発現レベルの平均値及び標準偏差によって表す。 In measuring any gene expression level, the template is

That is, it is represented by the mean and standard deviation of cyclin G2 expression levels and preferably Sharp1 expression levels in a population or data set.

  Cyclin G2 and Sharp 1 in two cell lines representing non-invasive and metastatic breast cancer, BT20 (ATCC # HTB-19) and MDA-MB-436 (ATCC # HTB-130), or other representative high standards The expression levels of the expression control and the low standard expression control are preferably added to the population values of the template.

Standard Expression Controls Standard expression controls include non-invasive and metastatic breast cancer samples or cell lines such as cyclin G2 alone in BT20 (ATCC # HTB-19) and MDA-MB-436 (ATCC # HTB-130). Or expression levels in combination with Sharp1 or other representative high and low levels of cyclin G2 alone or in combination with Sharp1.

Signature score (or expression score)
The signature score quantifies the difference between cyclin G2 expression values in the unknown sample compared to the template and preferably also Sharp1 expression values.

［式中、
サイクリンＧ２を単独で用いる場合にはＫ＝１であり、サイクリンＧ２及びＳｈａｒｐ１の両方を用いる場合にはＫ＝２であり、

は未知試料ｉ中のサイクリンＧ２又はＳｈａｒｐ１の発現レベルであり、

はそれぞれ、病歴が知られている母集団におけるサイクリンＧ２及び／又はＳｈａｒｐ１の発現レベルの推定平均値及び推定標準偏差値を表す］
と定義される。 The signature score is generally

[In the formula,
K = 1 if cyclin G2 is used alone, K = 2 if both cyclin G2 and Sharp1 are used,

Is the expression level of cyclin G2 or Sharp1 in unknown sample i,

Respectively represent estimated mean and estimated standard deviation values of expression levels of cyclin G2 and / or Sharp1 in a population with a known medical history]
It is defined as

＝サイクリンＧ２及びＳｈａｒｐ１の組合せのシグネチャースコア
［式中、

は未知試料ｉ中のＳｈａｒｐ１及びサイクリンＧ２の発現レベルであり、

はそのテンプレートを定義する］。 For cyclin G2 and Sharp1 expression measured in combination:

= Signature score of the combination of cyclin G2 and Sharp 1 [wherein

Is the expression level of Sharp1 and cyclin G2 in unknown sample i,

Defines its template].

［式中、

は未知試料ｉ中のサイクリンＧ２の発現レベルであり、

はそのテンプレートを定義する］
として計算される。 If the minimal signature template is obtained by measuring the expression level of cyclin G2 alone, the signature score is:

[In the formula,

Is the expression level of cyclin G2 in unknown sample i,

Defines the template]
Calculated as

Minimal Signature Minimal signature height is defined as a signature (expression) score higher than 0.
Minimal signature low is defined as a signature (expression) score lower than 0.

Relapse Relapse is defined as the onset of metastases associated with breast cancer (more commonly, metastases to the lung or bone) or breast cancer recurrence within 12 years of primary tumor surgery.

Control Assay Control-An "assay control" known to the person skilled in the art assesses the reliability of the criteria and taking of a signal upon which the assay produces consistent results. For example, a positive "assay control" for PCR is a known nucleic acid mixture in which PCR using primers is expected to result in the amplification of DNA fragments of the expected length.
Internal Expression Control-This term is generally used to indicate a housekeeping gene expression control.

[Detailed description]
The present invention is based on experimental evidence that the mutant allele of p53 cooperates with TGFβ to sustain its pro-invasive response and malignancy response. In fact, mutant p53 expression is required for in vitro invasion and in vivo metastatic spread, which reveals a previously unconfirmed link between these two pathways in breast cancer progression .

  The pro-invasion pathway activated by TGFβ in the manner of mutant p53 involves downregulation of cyclin G2 and Sharp1 genes. Lower expression levels of these genes correlate with breast cancer pro-invasive behavior and thus higher risk of cancer recurrence.

  The present invention has prediction power comparable to more complex gene set predictors, with cyclin G2 alone or cyclin G2 together with Sharp1 (cyclin G2 and Sharp1 in the following referred to as minimal signature (MS)) Indicates that. Due to the small number of genes involved in this evaluation, the invention can be practiced with commonly used techniques and simple PCR equipment.

  A correlation between minimal signature and breast cancer recurrence or metastatic spread was confirmed by statistical analysis on several breast cancer data sets using the expression levels of these two genes, but in one database, cyclin G2 alone caused cancer recurrence Indicated to predict.

  The method preferably comprises expression levels of cyclin G2 (gene ID = 901) from a plurality, preferably at least 50 to 100, of tumor patients whose clinical follow-up is known or available breast cancer patient datasets. Based on a large amount of minimal signature template used in combination with the expression level of Sharp1 (Gene ID = 79365).

  The present invention assesses the risk of recurrence of breast cancer patients, including detecting cyclin G2 (gene ID = 901) gene expression levels in unknown samples alone or in combination with Sharp 1 (gene ID = 79365) gene expression levels. Disclose the method.

［式中、
サイクリンＧ２を単独で用いる場合にはＫ＝１であり、サイクリンＧ２及びＳｈａｒｐ１の両方を用いる場合にはＫ＝２であり、

は未知試料ｉ中のサイクリンＧ２又はＳｈａｒｐ１の発現レベルであり、

はそれぞれ、病歴が知られている母集団におけるサイクリンＧ２及び／又はＳｈａｒｐ１の発現レベルの推定平均値及び推定標準偏差値を表す］
と定義される）を計算するステップと
（ｃ）未知試料を、当該シグネチャースコアが０より低い場合には最小シグネチャー低群に又は当該シグネチャースコアが０より高い場合には最小シグネチャー高群に分類する（低群への割り付けは高い再発リスクと相関する）ステップと
を含む、乳癌患者の「癌再発」リスクを評価するための方法を含む。 The present invention preferably
(A) detecting the gene expression level of cyclin G2 (gene ID = 901) in a sample from a breast cancer patient, preferably in combination with the gene expression level of Sharp (one or more) (gene ID = 79365) ( (A) measuring and obtaining a signal related to marker gene expression) and (b) a signature score of cyclin G2 alone in an unknown sample, or preferably a signature score of both cyclin G2 and Sharp 1 (the said signature score is

[In the formula,
K = 1 if cyclin G2 is used alone, K = 2 if both cyclin G2 and Sharp1 are used,

Is the expression level of cyclin G2 or Sharp1 in unknown sample i,

Respectively represent estimated mean and estimated standard deviation values of expression levels of cyclin G2 and / or Sharp1 in a population with a known medical history]
And (c) classify the unknown sample into the minimum signature low group if the signature score is lower than 0 or the minimum signature height group if the signature score is higher than 0. And (c) assigning to the low group correlates with high risk of relapse, including methods for assessing the risk of "cancer recurrence" in breast cancer patients.

  The sample can be a breast cancer biopsy or lymph node and tissue section or nucleic acid, preferably mRNA or cDNA isolated from such a sample.

  The high predictive power of the method of the invention of measuring cyclin G2 (gene ID = 901) alone or preferably in combination with Sharp 1 is a signature of only two genes which is superior to the more than 400 genes it is controlled by TGFβ. It is particularly surprising that all that is and the signatures already presented do not contain either of the two genes according to the present invention, and that their prognostic use for breast cancer recurrence is described for the first time herein. It is.

  Minimal signature templates are prepared by collecting gene expression data (ie cyclin G2 and preferably also Sharp1) from clinical data and a population of patients known to have a 5- 12 year survival period.

  Preferably, detection of one or preferably two marker genes in an unknown sample is a control for high expression level standards of each gene (control high Cyclin G2 and control high Sharp 1) and a control for low expression level standards (control The same reagent is used at the same time in control low cyclin G2 and control low Sharp1).

  Standard expression control high and low, respectively, can be obtained from known patients or from cell lines representative of non-invasive or metastatic breast cancer (eg, BT20 or MDA-MB-436, respectively). BT20 (ATCC # HTB-19) and MDA-MB-436 (ATCC # HTB-130) are two different breast cancer cell lines that represent non-invasive and metastatic breast cancer, respectively. BT20 expresses high levels of both genes and, conversely, in MDA-MB-436, Sharp1 and cyclin G2 are downregulated. Thus, these two cell lines can provide accessible high (BT20) and low (MDA-MB-436) standard expression controls for the proposed method.

  Furthermore, at least one internal expression control for normalization is measured in the same reaction.

  The choice of internal expression control depends on the experimental technique used to observe expression levels. Normalization of expression data can be done computationally (as scaling to the average expression level of all genes or quantile normalization) when using microarrays or nucleic acid based molecular techniques ie PCR or Northern blot Case can be based on the expression level of the internal control. For this purpose, for example, housekeeping genes commonly used for PCR are selected among constitutively expressed GAPDH, β-actin and the like. In the case of methods based on immunodetection, the internal control is preferably selected among Lamin B or GAPDH immunoreactivity. Also, additional assay controls known by the person skilled in the art are preferably included in the method to assess the reliability of steps a) and b), providing a reliable basis for the assay to produce consistent results. Provide a control. For example, a positive assay control for PCR is a known nucleic acid mixture where PCR using primers is expected to result in the amplification of DNA fragments of the expected length.

  Measurement of cyclin G2 and / or Sharp1 gene expression levels may be by any known state of the art methods, for example by molecular means based on molecular selection (ie selective amplification or hybridization) and / or by immunological means evaluate.

  After molecular selection (ie selection by sequence specific hybridization with a sequence specific probe or primer for cyclin G2 and / or Sharp 1), usually based on molecular weight, targeted and / or amplified polynucleotide molecules Separation step followed by quantification (eg by densitometry or visual inspection), then data normalization using any modern calculation method (eg linear scaling or non-linear normalization, and preferably comparison with standard expression control Do).

  Preferably, comparison of sample values to the minimum signature template is performed by calculation of the signature score.

  However, more generally, the present invention provides that if the expression level of the cyclin G2 gene in the sample alone or preferably in combination with the expression level of the Sharp1 gene manifests a signature score lower than 0, Based on the definition that (milk) cancer relapse risk shows increased signs.

  Preferably, statistical analysis is used to compare and / or distinguish individuals having one phenotype (eg, an unknown sample) with other individuals having a second phenotype (eg, a minimal signature template). Preferably, this is performed by software.

  Thus, according to a preferred embodiment, the method of the invention searches stored templates and numerically sets the sample's signature score by the expression level signal (s) of the marker (s) And b) implemented by software executed on a computer, and assigning unknown samples to high or low minimum signature groups (as specified in step c) above).

More preferably, the analysis of the signal (expression data) obtained (in accordance with step a) above is performed by the following additional steps: Data quality control based on assay controls,
Normalization of different data by technology, by techniques used to quantify gene expression levels
Preferably, it is performed by rescaling data based on standard expression controls, for example by linear or non-linear scaling.

  After suitably analyzing the signal, the template is searched, the signature score of the sample is calculated, and the unknown sample is assigned to said minimum signature high group or low group (as defined in step c)).

  If the signature template is stored on a computer or computer readable medium and the software is used for prognostic correlation signatures, the signature template is compared to the signature score from the sample. This, in other words, the expression level of one or both of the two marker genes in the sample, suitably and preferably analyzed, from the patient or alternatively or additionally, non-invasive and metastatic breast cancer Measured from a pool of samples from patients of known prognosis (ie a pool of numerically suitable samples, usually composed of at least 50 to 100 samples), including samples from representative cell lines It is meant to compare with the distribution of expression levels of the same gene in the minimal signature.

  Next, if the expression level of one or both of the two marker genes determines a signature score higher than 0, the unknown sample is classified as having a good prognosis for cancer recurrence. Conversely, unknown samples with a signature score lower than 0 are classified by the software as samples from patients with poor prognosis.

  The method is preferably implemented by software, but is not limited to this embodiment. In practice, the assignment of high and low expression groups can be made by visually inspecting the absolute expression signal of the sample in the presence of controls known to the person skilled in the art, and with this and high and low signature templates (or It can also be carried out by visual comparison or numerical comparison with the standard expression control defined above.

  Preferably, in order to increase the sensitivity of the comparison, the signal associated with the expression level can be normalized, for example by using various techniques, such as the average expression level of a set of control genes.

  In various embodiments, marker expression levels are normalized by the mean or median expression levels of a set of control markers (internal expression controls are GAPDH or β-actin in the case of nucleic acid based assays, immune) In the case of science based assays GAPDH and Lamin B).

  In another specific embodiment, normalization is performed by marker level normalization. Expression level data can be converted in any convenient manner, but preferably the expression signals are log converted prior to performing normalization and comparison. The normalized value is then compared to the minimal signature template. The minimal signature template is the same from a suitable pool of tumor patients known to have clinical follow-up and from various breast cancer cell lines (eg, BT20 and MDA-MB-436, respectively) representing non-invasive and metastatic breast cancer. It consists of the normalized and / or transformed expression levels of the same marker gene collected using experimental techniques and protocols.

  As an example, if the markers are represented by probes in a microarray, the expression level of each marker is expressed across all genes represented on the microarray, including any non-marker (ie non-cyclin G2 and non-Sharp1) genes It can be normalized by the mean or median of the levels.

  As mentioned above, the measurement of expression levels can be carried out by any known method, and molecular means include, for example, PCR (standard PCR or real time PCR), Northern blot or microarray analysis.

  According to Northern blot, total RNA samples are separated by electrophoresis according to size, and hybridization with labeled probes specific for cyclin G2 and / or Sharp 1 is performed.

  PCR or RT-PCR involves the reverse transcription of RNA samples to cDNA as a preliminary step, and by standard sequence analysis using known available software, eg Primer 3 (http://primer3.sourceforge.net By using PCR primers identified from published cyclin G2 and Sharp 1 sequences.

  Preferred cyclin G2 and Sharp 1 forward and reverse primers for PCR based molecular methods of the invention are also shown in the following table, including PCR primers for amplification of preferred internal control genes.

In the case of quantitative PCR (Q-PCR), the following preferred primers are used.

  One of the most widely used methods of gene expression analysis is by (micro) arrays. As with other types of expression data measurements, statistical analysis of unknown samples is by collecting a suitable number (at least 50-100) of measurements from breast cancer patients whose clinical follow-up is known. , Preliminary evaluation of minimal signature templates of cyclin G2 (gene ID = 901) alone or preferably in combination with Sharp 1 (gene ID = 79365). These data, ie, the aforementioned minimum signature template, can be predefined and relevant information stored on a computer for subsequent sample analysis.

The methods of the invention were validated in the following breast cancer microarray datasets.

  Classification of Sharp1 and cyclin G2 co-expression score high or low into one of two groups is preferably to sum standardized expression levels of Sharp 1 and cyclin G2 to give a total score of average value 0 To carry out.

前記式中、

は、試料ｉ中のＳｈａｒｐ１及びサイクリンＧ２の発現レベルであり、

は、データセット全体にわたって計算されたＳｈａｒｐ１及びサイクリンＧ２の推定平均値及び推定標準偏差値であり、最小シグネチャーテンプレートを表す。 Tumors are classified as minimal signature low if the total score is negative and as minimal signature high if the total score is positive.

In the above formula,

Is the expression level of Sharp1 and cyclin G2 in sample i,

Is the estimated mean and estimated standard deviation of Sharp1 and cyclin G2 calculated over the entire data set and represents the minimal signature template.

  In the case of the NKI data set, the samples had to be divided into high and low groups based only on cyclin G2 data, which corresponded to the minimal requirements of the prognostic validity of the method. In this data set (295 tumors), only one cyclin G2 based stratification still predicts metastasis.

［式中、

は、未知試料ｉ中のサイクリンＧ２の発現レベルであり、

は、テンプレートを定義する］
に従って、サイクリンＧ２スコアが負の場合には最小シグネチャー低に、サイクリンＧ２スコアが正の場合には最小シグネチャー高に分類する。 In fact, if the expression level of cyclin G2 alone is used to define the minimal signature template, the tumor will calculate the following:

[In the formula,

Is the expression level of cyclin G2 in unknown sample i,

Defines the template]
According to, the classification is classified as minimal signature low when cyclin G2 score is negative, and minimal signature high when cyclin G2 score is positive.

  Thus, the risk of cancer recurrence is rated "high" for the minimal signature low expression group.

  The same analysis described briefly above and in more detail in the experimental part for validation of two markers can be performed on any new or different data set. Thus, according to a further embodiment, the present invention relates to a method of analyzing a breast cancer microarray data set using expression values of cyclin G2 alone or in combination with Sharp1.

  Notably, by applying the method to all of the data set, the prognostic method of the present invention is “high” group when tested using univariate Kaplan-Meier survival analysis. In groups expressing low level minimal signatures, which show a significantly higher probability of causing relapse compared to (P value was in the range of 0.02 to 3 E-05 depending on the data set), It proved to be a strong predictor of breast cancer recurrence.

  Interestingly, the minimal signature based on both cyclin G2 and Sharp1 expression levels functioned as well as the 70 gene profile described in van't Veer et al., 2002, in stratification of patients by clinical outcome .

  The advantage of using a minimal signature based on only two genes instead of 70 genes is clearly evident.

  A further advantage of the method of the present invention is that expression of cyclin G2 and Sharp1 is statistically correlated with any of the risk of distant metastasis to bone and lung and thus independent of secondary tumorigenic site It is.

  Furthermore, the simplest way in which this method can be practiced is PCR, which requires only minimal equipment such as a PCR thermocycler and a tank for DNA separation by gel electrophoresis, but the invention is limited to this embodiment All available, commonly used to measure gene expression levels when applied to the detection of expression levels of cyclin G2 alone or in combination with Sharp1 as a prognostic marker for breast cancer recurrence risk, but not Relate to the way.

Thus, the method of the present invention
Standard PCR technology,
Real-time PCR (or Q-PCR using Taq man or Sybr Green technology),
Microarrays, optionally combined with sequences specific to other genes,
Deep sequencing (t Hoen et al., 2008), optionally combined with sequences specific for other genes
Northern blot,
Immunohistochemical staining with available antibodies against cyclin G2 and / or Sharp 1
Gene expression levels can be measured for specific mRNA or protein products based on any one of the gene expression analysis techniques, such as immunoblot.

  According to a preferred measurement technique of expression level, quantitative PCR or reverse transcription mRNA PCR, the cyclin G2 detection reagent is a cyclin G2 specific for cyclin G2 which comprises a 13-mer oligonucleotide derived from SEQ ID NO: 1 or a complementary sequence thereof. Oligonucleotides.

  For immunodetection, preferably, anti-cyclin G2 specific antibodies are used alone or in combination with anti-Sharp1 specific antibodies.

  Thus, in summary, according to a preferred embodiment of this method which also comprises the detection of Sharp 1 expression levels, the specific detection reagent consists in an oligonucleotide comprising at least a 13-mer oligonucleotide derived from SEQ ID NO: 2 or its complementary sequence. It is selected from the group consisting of Sharp 1 specific oligonucleotides, or anti Sharp 1 specific antibodies.

  A further embodiment of the present invention relates to cyclin G2 and preferably further to Sharp 1 gene expression specific detection means, ie a cyclin comprising a 13-mer oligonucleotide derived from SEQ ID NO: 1 or a poly or oligonucleotide comprising at least a complementary sequence thereof. Risk of breast cancer recurrence of a breast cancer patient comprising a G2 specific oligonucleotide or probe, and preferably a Sharp 1 specific oligonucleotide present in a poly or oligonucleotide comprising at least a 13-mer oligonucleotide derived from SEQ ID NO: 2 or a complementary sequence thereof It is an evaluation kit.

  In a further embodiment of the present invention, the present invention is a kit for evaluating the expression of cyclin G2 alone or in combination with Sharp1 in a sample from a breast cancer patient, comprising at least a cyclin G2 specific reagent, preferably An oligonucleotide comprising at least a 13-mer derived from SEQ ID NO: 1 or a complementary sequence thereof; preferably an oligonucleotide further comprising a Sharp 1 specific reagent, preferably a 13-mer derived from SEQ ID NO: 2 or a complementary sequence thereof; The invention relates to a kit comprising instructions for analyzing an unknown sample, identifying criteria to assign the unknown sample measurement to a defined minimum signature high group or low group. According to one preferred embodiment, software for statistical analysis and comparison of expression data (sample signature score) with the minimal signature template defined above (assignment to minimal signature low group is cancer recurrence of breast cancer patients) Correlate with increased risk).

  This kit contains cyclin G2 and Sharp1 expression control high and low (ie cyclin G2 and Sharp1 expression values measured in cell lines BT20 and MDA-MB-436 respectively) and dilution buffer or assay as standard expression controls A buffer can further be included.

  A specific reagent useful for each of the gene expression detection methods used may be a commercially available reagent or may be customized, as long as it is specific to cyclin G2 and / or Sharp1.

Antibodies, preferably purified polyclonal or monoclonal antibodies, or oligonucleotides can be labeled, preferably with fluorochromes, chemiluminescent labels or chromogens, and polynucleotides can be used in Northern blots, for example after labeling with ³² P .

  The specific antibody can be directly labeled or can be detected by using a secondary labeled antibody.

  The kit further comprises instructions for use in reporting criteria to assign each sample measurement to a high or low minimal signature (a low minimal signature correlates with an increased risk of cancer recurrence in breast cancer patients), or preferably Includes. Preferably, the calculations specified above are implemented in software.

  The kit can include reagents for detecting negative and positive samples and assay controls present in the positive sample, or internal expression controls, and optionally nucleic acid extraction reagents.

  According to one preferred embodiment, the PCR primer pair for cyclin G2 expression level detection is 5: CCTCCCCAGTGATCAAGAGTGC3 'for cyclin G2 (forward), 5' TCCCCTCTCCCCAAAAGT AGC3 'for cyclin G2 (reverse), and Sharp 1 (forward) It is 5 'GCATGAAACGAGACGACACC3', and Sharp 1 (reverse) is 5 'TCCCTCTCTCCCCAAAGTATC3'.

  Primers to be compared can be identified by known techniques.

Semi-quantitative PCR (RT-PCR) is typically polyA ⁺ RNA purified from total RNA extracted from the sample, using GAPDH sequences as internal expression controls, as known in the art. By retrotranscribing.
The expression level of each gene is obtained by densitometry analysis or visual inspection and comparisons with standard expression controls are performed to define low expression groups for cyclin G2 alone or for a combination of cyclin G2 and Sharp1.

  According to another embodiment, the kit comprises means for immunological detection of cyclin G2 and Sharp1 expression, such as specific antibodies and related controls.

  The results obtained by the method of the invention present a first stratification of recurrence risk for breast cancer patients.

  As mentioned above, the prognostic signs of cyclin G2 and Sharp 1 are one of the most important indicators for the physician, but physicians should be aware of age, tumor size, axillary lymph node status, tumor histotype, pathological grade and Prognostic assessment must be performed in conjunction with other known prognostic and predictive factors for breast cancer, such as hormone receptor status.

  In fact, as reported in more detail in the experimental part, Example 6, in Model 2, the tumor diameter, estrogen receptor's, was performed on a set of 187 tumors from the National Cancer Institute (Sotiriou et al., 2006). Commonly used in clinical practice, including status (ER positive vs. negative), lymph node status (positive vs. negative), tumor grade (grade 2 vs. grade 1 and grade 3 vs. grade 1) and treatment status (tamoxifen with vs. no) Multivariate Cox proportional hazards analysis of the other predictors that are found is very significant (p = 0.0054) for the smallest signature (Table 4).

  Thus, the minimal signature is a significant predictor of recurrence free survival, adding new prognostic information beyond that provided by standard clinical predictors. The minimal signature also adds prognostic value to multivariate models as well as any models calculated using any single clinical predictor. In fact, the residual deviation of the model obtained with a single clinical variable + minimal signature (eg, nodal status + minimal signature), of the model obtained with only one clinical variable There is a significant difference for each clinical predictor between the residual deviance.

  Furthermore, the method of the present invention provides prognostic indications for patients corresponding to more than 50% of breast cancer patients, who are secretly assigned unambiguously poor outcome or clearly good outcome by conventional prognostic markers. It is particularly useful.

  A particular relevance of this method is that it is effective for uncertain prognostic tumors (Ivshina et al., 2006) classified as moderate (grade 2) by the Nottingham scale, which represents the majority of tumors. It is to apply. When applied to grade 2 tumors from multiple independent data sets, the minimal signature stratified grade 2 samples into two groups with outcomes corresponding to grade 1 and grade 3, respectively.

  Thus, the division performed is applied to the stratification to more accurately classify breast cancer patients classified in grade 2 according to the Nottingham scale and possibly to assign them to different treatment categories or clinical trials. Is a preferred embodiment of the method of

EXPERIMENTAL PART Materials and Methods Cell Culture and Transfection H1299, and its derived cell lines expressing mutant p53 R175H, are as described in G. G. A gift from Blandino (Strano et al., J. Biol Chem 2002).

  H1299 non-small cell lung cancer cells were maintained in DMEM, 10% serum, 1 mM glutamine. TGFβ treatment was performed in DMEM 0.2% serum (TGFβ was provided by Peprotech). p53R175H H1299 cells express a stably transfected plasmid encoding a ponasterone derived cDNA for the mutant p53R175H allele. p53 expression was induced by incubating the cells with Ponasterone-A (Alexis, 3 mM) for 16 hours prior to treatment.

  MDA-MB-231 (ATCC # HTB-26) was maintained in a 1: 1 mixture of DMEM and F12 (DMEM / F12) supplemented with 10% serum, 2 mM glutamine.

  For TGFβ treatment, cells were serum starved for 24 hours and then treated with TGFβ1 (5 ng / ml) in DMEM / F12 without serum.

For transfection of siRNA (si is small interfering RNA), RNAi Max reagent (Invitrogen) was used to transfect the dsRNA oligo (10 pmol / cm ² ). A list of targeted sequences of siRNA and shRNA (sh is small hairpin RNA or short hairpin RNA) is shown in Table 1.

Generation of Stable Expression Cell Lines Short hairpin RNA (shRNA) expression constructs were generated by cloning annealed DNA oligonucleotides into pSUPER-retro-puro (OligoEngine). All plasmids were controlled by sequencing.

  For stable knockdown, retroviral particles were obtained by transfecting a plasmid expressing shRNA (pSuperRetro) and VSV envelope in 293 gp (as donated by M. Tripodi) using calcium phosphate. Two days after transfection, the supernatant was collected, filtered and used for MDA-MB-231 infection. After selection for puromycin resistance, transduced cells were verified for target protein downregulation.

Migration and Invasion Assay For wound-closure experiments, H1299 cells were plated in 6-well plates and cultured until confluence. Cells were scratched with p200 chip (time 0), transferred to low serum and treated as described above.

  For migration assays, transwell migration assays were performed in 24-well PET inserts (Falcon pore size 8.0 mm). For MDA-MB-231, cells were seeded in 10 cm dishes and transfected with siRNA and after 8 hours serum starved overnight. Then, 50,000 or 100,000 cells were plated in transwell inserts (at least 3 replicas for each sample) and either left untreated or treated with TGFβ1 (5 ng / ml). For H1299, cells were seeded in 10% serum in transwells, then changed to 0.2% serum. For both cell lines, cells at the top of the transwell were swabbed off, migrating cells were fixed in 4% PFA and stained with 0.5% Crystal Violet. The filters were photographed and the total number of cells was counted. Each experiment was independently repeated at least three times.

  For the Matrigel invasion assay shown in FIG. 2C, MDA-MB-231 and derived cell lines were diluted 1: 2 in DMEM / F12 with Matrigel Growth Factor Reduced (BD Biosciences). Resuspend in drops (100 ml).

In Vivo Metastasis Assay Mice were housed in a Specific Pathogen Free (SPF) animal facility and treated according to approved Institutional Standards (University of Padova). For xenograft studies of breast cancer metastasis, the mammary fat bodies of SCID female mice adapted to age 5 to 7 weeks with shGFP- or shp53-MDA-MB-231 cells (cell count 1 x 10 ⁶ / mouse). Was injected unilaterally. Six weeks later, mice were sacrificed and examined for the presence of lymph node metastases. Macroscopic metastases to other organs (liver, lung, peritoneum) were rare. Tumor growth at the injection site was observed by repeated caliper measurements. For the lung colony formation assay, cells were resuspended in 100 ml PBS and inoculated into the tail vein of SCID mice. Four weeks later, animals were sacrificed and lungs removed for subsequent histological analysis.

Histology and Immunohistochemistry Tissues for histological examination were fixed in 4% buffered formalin, dehydrated and paraffin embedded by standard methods.

  For the experiments shown in FIGS. 2G-I, serial sections of the lungs are cut 150 mm apart from each other and first stained with hematoxylin and Eosin (H & E) and then for human cytokeratin expression. Were treated with anti-human cytokeratin (Cytokeratin) mouse monoclonal antibody (clone MNF116) (Dako). Immunohistochemical staining was performed using an indirect immunoperoxidase staining technique (Bond Polymer Refine Detection, Vision BioSystems, UK).

  We quantified cytokeratin positive areas in 5 serial sections per lung. Areas covered by tumor cells were measured using ImageJ software (NIH) from four non-overlapping fields per section (covering 50-80% of each section).

Antibodies and Western Blotting Western blot analysis was performed as described previously (Piccolo et al., 1999). Briefly, proteins were separated in 10% NuPage® gel (Invitrogen) and transferred to Immobilon P® membrane (Millipore). Chemiluminescence was visualized using Supersignal West-pico® and -duraHRP substrate (Pierce). Anti-human p53 DO-1 monoclonal antibody and anti-lamin polyclonal antibody were purchased from Santa Cruz Biotechnology. Anti-phospho-Smad3 polyclonal antibody was purchased from Cell Signaling.

Northern Blotting Total RNA was extracted by Trizol (Invitrogen) from cells seeded in 6 cm dishes. 10 mg of total RNA per sample is separated by loading in 6% formaldehyde / 1% agarose gel, blotted by upward capillary transfer on GeneScreen Plus (PerkinElmer) and UV crosslinked The The membranes were prehybridized for 5 hours at 42 ° C. with ULTRAhyb-Oligo solution (Ambion) and hybridized with a ³² P-labeled DNA probe at 42 ° C. overnight. Membranes were washed at 68 ° C. with 2 × SSC / 0.5% SDS solution and exposed to autoradiography. All probes were obtained by random primer amplification. Sharp1, cyclin G2 and follistatin probe templates were obtained from RZPD EST (HU3_p983B0120D, HU3_p983D0140D2 and RZPD EST HU3_p983D0113D2, respectively). GPR87 and ADAMTS9 probes were obtained by cloning of RT-PCR products. All probes were verified by sequencing.

RT-PCR
Poly (A) ⁺ -RNA was retrotranscribed with M-MLV Reverse Transcriptase (Invitrogen) and oligo-d (T) primers after total RNA purification with Trizol (Invitrogen). In the case of standard RT-PCR, 2 μl of each cDNA sample is aliquoted into PCR tubes and then master PCR mix for EXTaq (Finnzymes) is added. The cycling conditions are 30 seconds at 94 ° C., 30 seconds at 55 ° C., 60 seconds at 72 ° C. (Cordenonsi et al., 2003). A list of all PCR primers is shown in Table 2.

  Q-PCR for cyclin G2 and GAPDH was performed by using 7500 Real-Time PCR System (Applied Biosystems) with DyNAmo HS SYBR Green (Finnzymes).

Microarray analysis MDA shGFP and shp53 cells were serum starved for 24 h and either left untreated or treated with TGFβ1 (5 ng / ml, 3 h) in DMEM / F12 without serum. Four replicas were prepared for each of the four conditions (untreated shGFP, TGFβ-treated shGFP, untreated shp53, TGFβ-treated shp53) (total of 16 samples). Total RNA was extracted with Trizol (Invitrogen) according to the manufacturer's instructions. Sample preparation for microarray hybridization was performed as described in the Affymetrix GeneChip® Expression Analysis Technical Manual. Briefly, 15 μg of total RNA was used to generate double stranded cDNA (Invitrogen). The synthesis of biotin labeled cRNA was performed using the BioarrayTM High YieldTM RNA Transcriptional Labeling Kit (ENZO Biochem, New York, NY). The length of cRNA fragmentation was confirmed using an Agilent 2100 Bioanalyzer (Agilent Technologies). Four biological mRNA replicates for each group were hybridized on Affymetrix GeneChip® Human Genome HG-U133 Plus 2.0 Array.

  All data analysis was performed on R using the Bioconductor library and R statistical package (http://www.r-project.org./, R Development Core Team, 2008). Specifically, the BioConductor packages AffyQCReport and AffyPLM were used for standard Affymetrix quality control procedures. Probe level signals were converted to expression values using the robust multi-array averaging procedure rma (Irizarry et al., 2003). In RMA, PM values were background corrected, normalized using quantile normalization, and expression criteria were calculated using median polish aggregation. RMA data with a standard deviation (e.g. 0.2) lower than the average standard deviation of all log signals in the whole array was filtered out. The filtered data set yielded 22644 probe sets used for further analysis. Differentially expressed genes were identified using microarray significance analysis samr (Tusher et al., 2001). SAM is a statistical method to detect significant genes in microarrays while controlling False Discovery Rate (FDR). SAM uses overlapping permutations of the data to determine whether the expression level of any gene is significantly related to the physiological condition, and the significant difference is the q-value, ie, the genes were "differentially expressed". It is quantified by the lowest false positive rate considered (Storey, 2002).

Identification of TGFβ Target Genes In order to identify genes whose expression is improved by TGFβ, we compare the expression profile of TGFβ treated MDA-MB-231 cells (shGFP or shp53) with their untreated controls And those transcripts with q values ≦ 0.1 were selected. This selection was further refined by setting the lower limit (or decrease) in TGFβ induction factor to 1.5. Using this combination filter, we are able to identify 447 differentially regulated genes between untreated MDA-MB-231 and TGFβ-treated MDA-MB-231 samples. there were. The differentially expressed genes are DAVID (http://david.abcc.ncifcrf.gov/), Kyoto Gene Genome Encyclopedia (KEGG, http://www.genome.jp/kegg/) and NCBI Gene Database ( Functionally classified according to NCBI, http://www.ncbi.nlm.nih.gov/sites/entrez?db=gene). Of the 292 genes associated with known functions, 147 genes were reported to be involved in cell motility, invasion processes and metastasis. A gene that is regulated by TGFβ1 in a mutant p53-dependent manner was confirmed to show significant control by TGFβ in shGFP but not in p53-deficient cells (q value ≦ 0.1, see FIG. 3B) . The resulting 5 genes were verified by Northern blot analysis.

Example 1
Effect of mutant p53 on cellular response to TGFβ.
We sought to probe the effects of mutant p53 on cellular responses to TGFβ. To achieve this goal, we used p53 deficient H1299 cells stably reconstituted with an inducible expression vector encoding the hotspot p53R175H mutant allele. As judged by P-Smad3 activation, this cell line possessed similar responsiveness to TGFβ as compared to the parental H1299 (FIG. 1A).

  Treatment of H1299 cells with p53R175H with TGFβ causes the cells to discard their cubic epithelial shape and display a more mesenchymal phenotype characterized by multiple dynamic projections such as filopodia and lamellipodia As it was acquired, it caused a marked morphological change (FIG. 1B). These were not present in either the parental or wild type p53 reconstituted cells (FIG. 1B and data not shown). To investigate whether expression of mutant p53 confers more chemotaxis to cells that receive TGFβ, we used a wound assay. In a wound assay, cells are induced to break cell-cell contacts, polarize, and migrate to a wound created by scratching a confluent culture with a pipette tip. After 30 hours of TGFβ treatment, parental (p53 deficient) H1299 cells did not migrate well, but p53R175H expressing cells almost completely infiltrated the wound (FIG. 1C). To attribute this effect to cell migration rather than growth bias, we observed BrdU uptake and found that there was no difference between TGFβ-treated control or mutant p53-expressing cells (data shows Not shown). As an independent measure of cell motility, we examined the behavior of parental H1299 cells, wild-type or mutant p53-reconstituted H1299 cells in a transwell-migration assay. FIG. 1D shows that expression of mutant p53 is consistent with gain of the TGFβ pro-migratory response, but wild-type p53 is not.

  These data link acquisition of mutant p53 to a phenotype in which TGFβ induced epithelial plasticity and migration, expression is important for TGFβ infiltration (Gupta and Massague, 2006).

Example 2
In vitro, mutant p53 and TGFβ jointly control the cell shape and invasiveness of breast cancer cells.
To clarify the practical requirement of increased epithelial plasticity and migration in metastatic cancer cells with endogenous mutant p53, we have established in MDA-MB-231 cells, an established model of invasive breast cancer. Stably knocked down the endogenous mutant p53 (p53R280K) (Arteaga et al., 1993; Bandyopadhyay et al., 1999; Deckers et al., 2006; Padua et al., 2008). Cells were transduced with a retroviral vector expressing shGFP (control) or shRNA targeting p53 (shp53) (see Table 1), followed by drug selection to accumulate positive transfectants. By immunoblot expression, shp53 expression reduced endogenous levels of mutant p53 protein by> 90% (FIG. 2A). In the transwell-migration assay, TGFβ triggered a strong pro-migratory response in control MDA-MB-231 cells. Remarkably, this response was abolished in mutant p53 deficient cells (FIG. 2B). Similar results were obtained upon transient depletion of p53 with two independent anti-p53 siRNA sequences (data not shown). When embedded in a drop of Matrigel, MDA-MB-231 cells show TGF.beta.-dependent scattering, extracellular matrix degradation and migration (FIGS. 2C and 2D), which reproduces in vivo invasiveness (FIG. 2C and 2D). Albini, 1998). We found that these activities require expression of mutant p53. These data suggest that, at least in vitro, mutant p53 and TGFβ cooperate to control the cell shape and invasiveness of breast cancer cells.

Example 3
Mutant p53 expression plays a pivotal role in directing TGFβ responsiveness in an efficient metastatic spread direction in vivo.
Several lines of evidence indicate that metastatic spread of MDA-MB-231 cells in vivo is under the control of autocrine TGFβ (Arteaga et al., 1993; Bandyopadhyay et al., 1999; Deckers et al., 2006). Padua et al., 2008). In order to test whether the mutated p53 has an importance related to the malignant behavior promoted by TGFβ in vivo, we used shGFP- or shp53-MDA- to the mammary fat bodies of immunodeficient mice. MB-231 cells were injected. The two cell populations grew at similar rates in vitro (data not shown) and formed primary tumors at similar rates and sizes in vivo (FIG. 2E). This indicates that high levels of mutant p53 in MDA-MB-231 cells are not essential for growth or primary tumorigenesis. Six weeks after transplantation, mice were sacrificed and examined for metastatic lesions.

  Orthotopically injected MDA-MB-231 has very few lung metastases but effectively metastasizes to the lymph nodes. To quantify metastatic spread, we observed colonization of the contralateral lymph node, a readout of systemic disease in human breast cancer (Singletary et al., 2006). Particularly noteworthy, in mice injected with shGFP cells, only one mouse out of 22 showed a negative score for lymph node metastasis, whereas mice carrying shp53-deficient tumors Since, in 10 of 22 mice remained metastatic free (Figure 2F), suppression of mutant p53 expression significantly reduced the number of lymph node metastases when compared to control cells .

  To corroborate these results in vivo that associate mutant p53 with invasiveness, we injected nude mice with a control and shp53-MDA-MB-231 intravenously. Using two independent clones, we show that depletion of mutant p53 has a marked effect on colonization of the lungs and a marked reduction in metastatic nodules in number and size (Figure 2G-2I) were found. Thus, mutant p53 expression plays a crucial role in directing TGFβ responsiveness in the direction of efficient metastatic spread.

Example 4
Identification of gene sets co-regulated by mutant p53 and TGFβ.
Next, we sought to scrutinize specific gene expression programs in which mutant p35 and TGFβ control invasion and metastasis. To identify this set of genes, we compared the TGFβ transcriptome profiles of control and mutant p53 deficient MDA-MB-231 cells. We found that TGFβ could regulate more than 400 genes. The majority of them were expressed independently of the presence of mutant p53.

  Some of the mutated p53 independent targets were already described in the literature as direct Smad targets such as PAI1 / SERPINE1, JunB and Smad7 (Massague and Gomis, 2006). In addition, genes related to lung or bone specific metastasis (ANGPTL4, NEDD9, IL11 and CTGF) (Padua et al., 2008) have already been linked to the general epithelial "TGF beta response classifier", including Many genes that have been found were also found. The successful identification of these targets validated our approach to identifying novel genes that may play an important role in TGFβ induced malignancy. Interestingly, we focused on 147 genes already considered to be involved in cell motility, invasion or metastasis (FIG. 3A and data not shown).

  However, TGFβ requires the presence of mutant p53 to exert its pro-metastatic function. Therefore, we focused on the much smaller set of genes co-regulated by mutant p35 and TGFβ. Notably, only five genes were required for this, Sharp1 / DEC2 / BHLHB3 / BHLHE41, cyclin G2 / CCNG2, ADAMTS9, follistatin and GPR87 (see FIGS. 3B and 3C). In particular, we focused on two candidate metastasis suppressors, Sharp1 and CylinG2, which are negatively regulated by TGFβ via mutant p53 (FIG. 3D). Sharp 1 is an inhibitory basic helix loop helix similar to ID-proteins (ie in the MyoD inhibition assay (Li et al., 2003), but its biological role is otherwise largely unknown . Cyclin G2 is considered an atypical "inhibitory" cyclin, but may also influence the dynamics of the microtubule cytoskeleton. Interestingly, cyclin G2 is inherited asymmetrically by association with centrosomes that surround the centriole during cell division (Arachchige Don et al., 2006).

Example 5
In vitro biological validation of identified gene sets.
To functionally validate these genes as effectors of the mutant p53 / TGFβ pathway, we performed epistasis experiments and Sharp1 or Cyclin G2 deficiency can rescue TGFβ-induced migration in p53 deficient cells It was tested. As shown in FIG. 3E, siRNA-mediated knockdown of Sharp1 or cyclin G2 restores TGF.beta.-dependent promigration activity in shp53 MDA-MB-231 (compare lanes 3 and 4 with lane 2 in FIG. 3E) . Thus, these molecules antagonize the TGFβ pre-invasive response to act as a metastasis suppressor. After identifying the genes essential to antagonize invasion behavior in vitro, we sought to elucidate their clinical relevance as metastasis suppressors. Recent transcriptome profiling of human primary tumors has identified a set of genes or "signatures" that predicts high risk of metastasis and low disease-free survival (Fan et al., 2006; van't Veer et al., 2002) . If the detection of Sharp1 and cyclin G2 in primary tumors is biologically relevant, reduced expression of these genes may be expected to be associated with poor clinical outcome. Surprisingly, Sharp1 and cyclin G2 are not included in the known signatures of breast cancer metastasis, ie the 70 gene signature, recurrence score etc. (Fan et al., 2006).

Example 6
Verification of prognostic value of identified gene sets by statistical analysis and comparison with other gene sets.
Breast Cancer Datasets To evaluate the prognostic value of Sharp1 and cyclin G2, we collected 6 different data sets (Table 3). For each data set, we performed survival analysis to test whether the minimal signature could classify patients into clinically different groups. Each data set was processed independently of one another to maintain the initial differences between the different studies (eg, patient cohort, microarray type, sample processing protocol).

  To assess the prognostic value of Sharp1 and cyclin G2 (minimal signature, MS), we utilized up to 900 primary breast cancers for survival and distant recurrence using available gene expression data sets Together with relevant clinical data, including

  We combined breast cancer gene expression data sets with clinical information Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/GEO/), Stanford microarray database (http: //genome-www5.stanford Or from the author's personal web page (http://microarray-pubs.stanford.edu/wound_NKI/explore.html).

  Table 3 presents a complete list of data sets and their sources. Raw data (eg, CEL files) were available for all samples, except for the EMC, MSK and NKI studies. Detailed clinical information could be obtained for all analysis samples. The data set included both Affymetrix and a dual channel cDNA microarray platform. As all Affymetrix data were derived from the same HG-U133A platform, there was no need for a method to map probe sets across the various generations of Affymetrix gene chip arrays. If a CEL file is available, use the RMA algorithm to generate expression values from the intensity signal, background correct the values, normalize using quantile normalization, and calculate expression criteria using median polish aggregation did. For EMC, MSK and NKI studies, data were used as downloaded. Specifically, in the EMC and MSK data sets, expression values were calculated using the Affymetrix MAS 5.0 algorithm. In the Affymetrix HG-U133A array, cyclin G2 is represented by three probe sets (202769_at, 202770_s_at and 211559_s_at), and Sharp 1 is represented by probe set 221530_s_at only.

  The Agilent Rosetta Inpharmatics array used for the NKI data set has a single probe for cyclin G2 but no probe for Sharp1.

Minimal Signature Classification In order to identify two sample groups by high or low co-expression score of Sharp1 and cyclin G2, we aggregate normalized expression levels of Sharp1 and cyclin G2 and sum score of mean 0 Defined the classification rules.

前記式中、

は、試料ｉ中のＳｈａｒｐ１及びサイクリンＧ２の発現レベルであり、

は、データセット全体にわたって計算されたＳｈａｒｐ１及びサイクリンＧ２の推定平均値及び推定標準偏差値である。 The tumors were then classified as minimal signature low if the total score is negative and as minimal signature high if the total score is positive.

In the above formula,

Is the expression level of Sharp1 and cyclin G2 in sample i,

Is the estimated mean and estimated standard deviation of Sharp1 and cyclin G2 calculated across the data set.

  For the Stockholm, NCI and Uppsala studies, this classification was applied based on the expression values obtained from RMA. On the other hand, for EMC and MSK, expression values are used as downloaded. For the EMC data set, expression data is log2-transformed.

  For the NKI data set, samples had to be classified into high and low groups based only on cyclin G2 data.

  In order to determine the appropriate threshold of cyclin G2 expression levels, we use clinical parameters to achieve good clinical outcome, ie, node-negative patients who remain metastasis free after at least 5 years of follow-up. Was quantified (van't Veer et al., 2002). As approximately 31% of the samples (92 out of 295 tumors) met these criteria, the 69th percentile of cyclin G2 expression values (ie, 0.078) could be tumor in the high or low group Used as a cutoff for classification. If the cyclin G2 expression level of a given sample was higher than the 69th percentile of cyclin G2 value, the sample was referred to as minimal signature high, otherwise it was referred to as minimal signature low. The rationale behind this choice is that approximately 31% of patients were expected to be classified as minimum signature height.

  Samples were also classified into minimal signature high groups and minimal signature low groups based on expression levels of Sharp1 and cyclin G2 using clustering methods without objective variables (Pollard, 2005).

  In particular, classification of MSK and EMC data sets uses aggregation type clustering with Euclidean distance and full or Ward connection criterion respectively, and NCI samples apply division type clustering (Diana) with Euclidean distance, Stockholm and Uppsala The k-means partitioning algorithm was used for the data set. The clustering method was not applied to NKI samples as only gene expression data for cyclin G2 is available.

  We compared the performance of the minimum signature and the 70 gene signature for all data sets analyzed. Since all datasets except NKI are from the Affymetrix array, we first map the 70 gene signature to the Affymetrix probe set, then in the Affymetrix U133A platform, which corresponds to the 48 unique EntrezGene IDs. We obtained a 70-gene poor prognostic signature map of NKI to 75 probe sets. Given this reduction in the number of genes that make up the signature, and in light of the fact that we used a different patient classification model, we have been generated based on the reduced number of gene lists It was verified whether the prognostic outcome of different models (ie clustering without objective variables) was similar to that of the van't Veer model based on all signatures. Therefore, we classified NKI samples using the 48 unique genes present on both the Affymetrix and Rosetta platforms and a classification model based on clustering without objective variables. Consistent with the results already reported by van't Veer et al. In 2002 and by Minn et al. In 2005, we use the clustering without objective variables to a reduced number of signatures as a classification indicator I found that it had little influence on my grades. Thus, samples in all other data sets were classified into two groups using this number of reduced 70 gene signatures and clustering without objective variables. In particular, in Stockholm's work, a cohesive hierarchical model based on Ward's algorithm (Ward, 1963) was used, and Uppsala and ECM's work was classified using PAM algorithm (Kaufman and Rousseeuw, 1990). Finally, for MSK studies, we used the classification presented by Minn et al. (2005).

Survival Analysis To assess the prognostic value of the minimal signature, we use the Kaplan-Meier method (Prentice, 1978) to identify patients according to whether they belong to high or low groups. The possibility of metastasis remaining (MSK and NKI), tumor recurrence remaining (Stockholm and NCI) and cancer disease remaining (Uppsala) was estimated. To confirm the outcome of these studies, survival curves are validated using Log-rank test or Mantel-Haenszel test (Harrington and Fleming, 1982), ie to support the survival of the smallest signature high. The null hypothesis with no difference was tested and compared against the one-sided alternative hypothesis. P values are calculated according to standard normal asymptotic distribution, corrected according to Bonferroni-Holm's sequential multiple test method (Dudoit, 2003), and the error rate per family ( Control family-wise error rate). All corrected p-values were significant at the significance level α = 0.05 when comparing the minimum signature high group and the minimum signature low group defined using the total score. Similar results are obtained with the same survival analysis repeated for the minimum signature high group and the minimum signature low group defined using the clustering method, with p values of 0.00026 for Stockholm, 0.00083 for NCI, for EMC It was 0.0251, 0.0025 for Uppsala, and 0.00887 for MSK.

  Finally, survival analysis was applied to a subset of samples assigned to high and low groups and classified as moderate (grade 2) by the Nottingham scale. Again, by controlling the error rate per family to α = 0.05, all null hypotheses were rejected. For the NCI data set, this analysis could not be performed as the recurrence free survival curve for grade 2 tumors was not statistically significantly different from the curve for poorly differentiated grade 3 tumors. Information on classification of tumors by the Nottingham scale is not valid in the MSK and EMC data sets.

Conclusion After defining two tumor groups with high and low levels of Sharp1 and CylcinG2 expression respectively in each data set (Figure 4), the test was performed using univariate Kaplan-Meier survival analysis. It should be found that groups expressing low level minimal signatures were found to have a significantly higher probability of relapse compared to the 'high' group (p values are 0 depending on the data set) 0.2-3 E-05).

  Interestingly, in stratification of patients by clinical outcome, MS showed a function comparable to the 70 gene profile (Figure 4).

  The expression of Sharp1 and cyclin G2 is synergistic with the predictive power of the minimal signature in these assays and is associated with both distant bone and lung metastatic risk (FIG. 5). Nevertheless, in patient datasets such as the NKI dataset (295 tumors) (Fan et al., 2006) where Sharp1 expression data is not valid, stratification based solely on cyclin G2 still predicts metastasis ( See Figure 6).

Multivariate analysis using the Cox proportional hazards model To further evaluate the prognostic value of the minimal signature, we used a multivariate data set of 187 tumors from the National Cancer Institute (Sotiriou et al., 2006). Cox proportional hazard analysis was conducted. In particular, the relapse risk for 187 tumors from the NCI study was examined by Cox proportional hazards regression modeling (Cox, 1972).

  Survival rates and minimal signature predictors and other predictors commonly used in clinical practice (tumor diameter, estrogen receptor status (ER positive vs. negative), lymph node status (positive negative tolerance), tumor grade (grade) The correlation with 2 vs grade 1 and grade 3 vs grade 1) and treatment status (including tamoxifen with vs. no) was specifically examined.

  We fitted a Cox proportional hazards regression model by first using only clinical variables (Model 1) and then adding a minimal signature predictor (Model 2). The results are shown in Tables 4 and 5. These tables show that the minimal signature is still a significant predictor of metastasis-free survival and thus has added new prognostic information beyond that provided by standard clinical predictors.

Table 4 Multivariate Analysis of Relapse Risk for NCI Datasets Using Cox Proportional Hazards Model In Model 1, tumor size and grade 2 (vs. grade 1) covariates are statistically significant at α = 0.05 It has a coefficient. However, when incorporating the minimal signature (Model 2), reassignment to the “low” group while keeping all other covariates constant, relapse hazards on average e ^0.706 = 2.026, significant Increase, ie add new prognostic information.

Model 1 Multivariate Analysis Using Only Clinical Variables Model 1 was obtained using n = 159 observations. The residual deviance (i.e.-partial log likelihood x 2) is equal to RD1 = 492.8774.

Model 2 Multivariate Analysis with Clinical Variables and Minimal Signatures Model 2 was obtained using n = 159 observations. The residual deviance (i.e.-partial log likelihood x 2) is equal to RD2 = 486.8369.

  Model 1 and Model 2 can be compared to assess whether the minimal signature adds additional prognostic information above clinical variables. In particular, it subtracts the residual deviance of model 1 (RD1 = 492.8774) from the residual deviance of model 2 (RD2 = 486.8369) and frees this difference (RD1-RD2 = 6.04043) It is obtained by testing against a power-of-one-squared distribution. Because this difference exceeds the 0.95 quantile of the one-degree-of-freedom distribution (p-value = 0101398), the minimal signature is a significant predictor of relapse-free survival and is above the standard clinical predictor Add new prognostic information.

  Furthermore, the minimal signature adds prognostic value not only to multivariate models, but also to any models created using any single clinical predictor. In fact, the residual deviation of the model obtained with a single clinical variable and a minimal signature (eg tumor diameter + minimal signature) and the residual of the model obtained with only one clinical variable The difference in deviance is significant for any clinical predictor.

  From the data, it was confirmed that the present invention provides an additional prognostic tool to assess the risk of metastasis, thus identifying patients for whom adjuvant therapy is effective.

  Furthermore, a preferred example is a tumor (Ivshina et al., 2006) classified as moderate (grade 2) by the Nottingham scale, which represents the majority of the tumor and has uncertain prognosis. When applied to grade 2 tumors from multiple independent data sets, the minimal signature divided these patients into two groups with outcomes corresponding to grade 1 and grade 3 respectively (Figure 7). This result was not obtained by all other more complex molecular methods and is therefore unique to the present invention.

References Albini, A. (1998). Tumor and endothelial cell invasion of basement membranes. The matrigel chemoinvasion assay as a tool for dissecting molecular mechanisms. Pathol Oncol Res 4, 230-241.
Arachchige Don, A. S. , Dallapiazza, R., et al. F. , Bennin, D. A. , Brake, T. , Cowan, C .; E. , And Horne, M. C. (2006). Cyclin G2 is a centrosome-associated nucleocytoplasmic shuttling protein that influences microtubule stability and induces a p53-dependent cell cycle arrest. Experimental cell research 312, 4181-4204.
Arteaga, C. L. , Hurd, S. D. , Winnier, A. R. , Johnson, M .; D. , Fendly, B. M. , And Forbes, J.A. T. (1993). Anti-transforming growth factor (TGF) -beta antibodies inhibit breast cancer cells Tumorigenicity and increase mouse spleen natural killer cell activity. Implications for a possible role of tumor cell / host TGF-beta interactions in human breast cancer progression. The Journal of clinical investigation 92, 2569-2576.
Bandyopadhyay, A. , Zhu, Y. , Cibull, M. L. , Bao, L. , Chen, C .; , And Sun, L. (1999). A soluble transforming growth factor beta type III receptor suppressive tumorigenicity and metastasis of human breast cancer MDA-MB-231 cells. Cancer research 59, 5041-5046.
Beenken, S. W. , Grizzle, W. E. , Crowe, D. R. , Conner, M. G. , Weiss, H. L. , Sellers, M. T. , Krontiras, H. , Urist, M. M. , And Bland, K. I. (2001). Molecular biomarkers for breast cancer prognosis: coexpression of c-erbB-2 and p53. Annals of surgery 233, 630-638.
Cordenonsi, M. , Dupont, S. , Maretto, S. , Insinga, A. , Imbriano, C. , And Piccolo, S. (2003). Links between tumor suppressors: p53 is required for TGF-beta gene responses by operating with Smads. Cell 113, 301-314.
Cox, D. R. (1972). Regression Models and Life Tables (with Discussion). Journal of the Royal Statistical Society, Series B-Statistical Methodology 34, 34.
Deckers, M. , Van Dinther, M. , Buijs, J., et al. , Que, I. , Lowik, C .; , Van der Pluijm, G., et al. , And ten Dijke, P .; (2006). The tumor suppressor Smad4 is required for transforming growth factor beta-induced epithelial to mesenchymal transition and bone metastasis of breast cancer cells. Cancer research 66, 2202-2209.
Dudoit, S. , Popper Shaffer. J. , Boldrick, J., et al. C. (2003). Multiple Hypothesis Testing in Microarray Experiments. Statistical Science 18, 71-103.
Fan, C .; , Oh, D. S. , Wessels, L .; , Weigelt, B. , Nuyten, D. S. , Nobel, A. B. , Van't Veer, L. J. , And Perou, C. M. (2006). Concordance among gene-expression-based predictors for breast cancer. The New England journal of medicine 355, 560-569.
Gupta, G. P. , And Massague, J.A. (2006). Cancer metastasis: building a framework. Cell 127, 679-695.
Harrington, D. P. , And Fleming, T. R. (1982). A class of rank test procedures for censored survival data. Biometrika 69, 4.
t Hoen, P .; A. , Ariyurek, Y. , Thygesen, H. H. , Vreugdenhil, E .. , Vossen, R. H. , De Menezes, R. X. , Boer, J.J. M. , Van Ommen, G., et al. J. , And den Dunnen, J.J. T. (2008). Deep sequencing-based expression analysis shows major advances in robustness, resolution and inter-lab portability over five microarray platforms. Nucleic acids research 36, e141.
Hartigan, J.J. A. , And Wong, M. A. (1979). A K-means clustering algorithm. Applied Statistics 28, 9.
Irizarry, R .; A. , Bolstad, B. M. , Collin, F .; , Cope, L. M. , Hobbs, B. , And Speed, T. P. (2003). Summaries of Affymetrix GeneChip probe level data. Nucleic Acids Res 31, e15.
Ivshina, A. V. , George, J. , Senko, O. , Mow, B. , Putti, T. C. , Smeds, J., et al. , Lindahl, T .; , Pawitan, Y. , Hall, P .; , Nordgren, H. , Et al. (2006). Genetic reclassification of histologic grade delineates new clinical subtypes of breast cancer. Cancer research 66, 10292-10301.
Li, Y. , Xie, M. , Song, X. , Gragen, S. , Sachdeva, K. , Wan, Y. , And Yan, B. (2003). DEC1 negatively regulates the expression of DEC2 through binding to the E-box in the proximal promoter. The Journal of biological chemistry 278, 16899-16907.
Miki, Y. , Swensen, J., et al. , Shattuck-Eidens, D .; , Futreal, P., et al. A. , Harshman, K. , Tavtigian, S. , Liu, Q. , Cochran, C. , Bennett, L .; M. , Ding, W. , Et al. (1994). A strong candidate for the breast and ovarian cancer susceptibility gene BRCA1. Science (New York, NY 266, 66-71.
Miller, L .; D. , Smeds, J., et al. , George, J. , Vega, V. B. , Vergara, L. , Ploner, A. , Pawitan, Y. , Hall, P .; , Klaar, S. , Liu, E. T. , Et al. (2005). An expression signature for p53 status in human breast cancer predicts mutation status, transcriptional effects, and patient survival. Proceedings of the National Academy of Sciences of the United States of America 102, 13550-13555.
Minn, A. J. , Gupta, G. P. , Siegel, P .; M. , Bos, P .; D. , Shu, W. , Giri, D. D. , Viale, A. , Olshen, A. B. , Gerald, W. L. , And Massague, J.A. (2005). Genes that median breast cancer metastasis to lung. Nature 436, 518-524.
Padua, D. , Zhang, X. H. , Wang, Q. , Nadal, C. , Gerald, W. L. Gomis, R., et al. R. , And Massague, J.A. (2008). TGFbeta primes breast tumors for lung metastasis seeding through angiopoietin-like 4. Cell 133, 66-77.
Pawitan, Y. , Bjohle, J., et al. , Amler, L., et al. , Borg, A. L. , Egyhazi, S. , Hall, P .; , Han, X. , Holmberg, L .; , Huang, F. , Klaar, S. , Et al. (2005). Gene expression profiling spares early breast cancer patients from adjuvant therapy: derived and validated in two population-based cohorts. Breast Cancer Res 7, R 953-964.
Piccolo, S. , Agius, E. , Leyns, L. , Bhattacharyya, S. , Grunz, H. , Bouwmeester, T. , And De Robertis, E. M. (1999). The head inducer Cerberus is a multifunctional antagonist of Nodal, BMP and Wnt signals. Nature 397, 707-710.
Pollard, K. S. , Van der Laan, M. J. (2005). Cluster Analysis of Genomic Data with Applications in R. U. C. Berkeley Division of Biostatistics Working Paper Series Working Paper 167.
Prentice, R. L. , Gloeckler, L. A. (1978). Regression Analysis of Grouped Survival Data with Applications to Breast Cancer Data. Biometrics 34, 57-67.
Singletary, S. E. , And Connolly, J.A. L. (2006). Breast cancer staging: Working with the sixth edition of the AJCC Cancer Staging Manual. CA: a cancer journal for clinicians 56, 37-47.
Sotiriou, C. , Wirapati, P .; , Loi, S. , Harris, A. , Fox, S. , Smeds, J., et al. , Nordgren, H. , Farmer, P .; , Praz, V. , Haibe-Kains, B., et al. , Et al. (2006). Gene expression profiling in breast cancer: understanding the molecular basis of histologic grade to improve prognosis. Journal of the National Cancer Institute 98, 262-272.
Storey, J. D. (2002). A direct approach to false discovery rates. Journal of the Royal Statistical Society Series B-Statistical Methodology 64, 479-498.
Tusher, V. G. , Tibshirani, R., et al. , And Chu, G.S. (2001). Significance analysis of microarrays applied to the ionizing radiation response. Proc Natl Acad Sci USA 98, 5116-5121.
van 't Veer, L. J. , Dai, H. , Van de Vijver, M. J. , He, Y. D. , Hart, A. A. , Mao, M. , Peterse, H. L. , Van der Kooy, K. , Marton, M. J. , Witteveen, A. T. , Et al. (2002). Gene expression profiling predicts clinical outcome of breast cancer. Nature 415, 530-536.
van de Vijver, M. J. , He, Y. D. , Van't Veer, L. J. , Dai, H. , Hart, A. A. , Voskuil, D. W. , Schreiber, G., et al. J. , Peterse, J.J. L. , Roberts, C .; , Marton, M. J. , Et al. (2002). A gene-expression signature as a predictor of survival in breast cancer. The New England journal of medicine 347, 1999-2009.
Wang, X. J. , Greenhalgh, D. A. , Jiang, A. , He, D. , Zhong, L. , Medina, D. , Brinkley, B .; R. , And Roop, D. R. (1998). Expression of a p53 mutant in the epidermis of transgenic mice accelerates chemical carcinogenesis. Oncogene 17, 35-45.
Ward, J. H. (1963). Hierarchical Grouping to optimize an objective function. Journal of American Statistical Association 301, 9.

Claims (21)

  A method for assessing the risk of recurrence of a breast cancer patient, comprising the step of detecting the gene expression level of cyclin G2 (gene ID = 901) alone or in combination with the gene expression level of Sharp 1 (gene ID = 79365) in a sample.   The method according to claim 1, wherein said detecting comprises measuring a signal and acquiring it. Calculating the signature score of cyclin G2 alone or preferably both cyclin G2 and Sharp 1 in an unknown sample, said signature score

[In the formula,
K = 1 if cyclin G2 is used alone, K = 2 if both cyclin G2 and Sharp1 are used,

Is the expression level of cyclin G2 or Sharp1 in unknown sample i,

Respectively represent estimated mean and estimated standard deviation values of cyclin G2 expression levels alone or in combination with Sharp1 expression levels in a breast cancer patient population of known history.
3. A method according to claim 1 or 2, defined as: and wherein a signature score of 0 or less indicates an increased risk of breast cancer recurrence.   The method according to any one of claims 1 to 3, wherein said detection is performed by molecular and / or immunological means.   The method according to any one of claims 1 to 4, wherein the molecular means is selected from the group consisting of PCR, microarray analysis, deep sequencing, Northern blot.   The method according to claim 5, wherein the PCR is real time PCR or quantitative PCR.   The method according to any one of claims 1 to 6, wherein the sample is a breast cancer biopsy material or a nucleic acid isolated from the breast cancer biopsy material. Quality control of acquired signals;
Normalizing the signal,
8. The method according to any one of claims 2-7, optionally further comprising rescaling the signal. i) Mean and standard deviation of Sharp1 and cyclin G2 expression values in a population of samples of known medical history

Defining the minimum signature template present in
ii) calculating a signature score according to claim 3 for cyclin G2 gene expression or cyclin G2 and Sharp1 gene expression in an unknown sample,
iii) Calculation

[In the formula,

Is the expression level of Sharp1 and cyclin G2 in the unknown sample,

Is the estimated mean value and estimated standard deviation value of Sharp1 and cyclin G2 calculated over the dataset consisting of samples of known history.]
Further classifying the unknown sample into a minimal signature low group if its signature score is negative or a minimal signature high group if its signature score is positive, to a minimal signature low group The method according to any one of claims 3 to 8, wherein the classification of is indicative of a high risk of cancer recurrence in breast cancer patients. At least the steps of acquiring a signal,
Quality control of acquired signals;
The method according to claim 8 or 9, wherein the step of normalizing the acquired signal is performed by computer-implemented software.   11. The method according to claim 10, wherein steps i) to iii) according to claim 9 are also implemented by computer implemented software.   Method of analyzing a breast cancer data set comprising gene expression data of cyclin G2 and / or Sharp1 comprising calculating the minimal signature template of i) according to claim 9 for gene expression data of cyclin G2 and preferably also Sharp1 .   Use of cyclin G2 (gene ID = 901) gene expression to evaluate cancer recurrence risk of breast cancer patients.   The use according to claim 13, further comprising the evaluation of Sharp1 gene expression (gene ID = 79365).   The use according to claim 13 or 14 for further division of breast tumors classified as moderate (grade 2) according to the Nottingham scale. The cyclin G2 gene expression is
i) A cyclin G2 specific oligonucleotide which is an oligonucleotide comprising at least a 13-mer oligonucleotide derived from SEQ ID NO: 1 or its complementary sequence,
ii) The use according to any one of claims 13-15, as measured by a detection reagent selected from the group consisting of anti-cyclin G2 specific antibodies. Said Sharp1 gene expression is
i) a Sharp 1 specific oligonucleotide which exists in an oligonucleotide containing at least a 13-mer oligonucleotide derived from SEQ ID NO: 2 or a complementary sequence thereof,
The use according to any one of claims 14 to 16, measured by a detection reagent selected from the group consisting of ii) anti-Sharp 1 specific antibodies. A kit for evaluating cyclin G2 expression alone or in combination with Sharp1 expression in a sample from a breast cancer patient and measuring the risk of cancer recurrence,
An oligonucleotide present in a cyclin G2-specific reagent, preferably an oligonucleotide comprising at least a 13-mer oligonucleotide derived from SEQ ID NO: 1 or a complementary sequence thereof
An oligonucleotide according to a Sharp 1 specific reagent, preferably an oligonucleotide comprising at least a 13-mer oligonucleotide derived from SEQ ID NO: 2 or a complementary sequence thereof
According to the calculation of i) to iii) according to claim 9, the signature score of the unknown sample is calculated, and the unknown sample is in the minimum signature low group if the signature score is negative or if the signature score is positive. A kit containing instructions for classifying into a minimum signature high group, and a kit showing classification into a minimum signature low group is an increased risk of cancer recurrence in breast cancer patients.   19. The kit of claim 18, wherein the instructions are included in the software.   The kit according to claim 18 or 19, further comprising cyclin G2 and Sharp 1 standard expression control high and low, expression value or nucleic acid sample as reference standard.   The kit according to claim 20, wherein the expression value or nucleic acid sample is derived from a non-metastatic breast cancer cell line and / or a high metastatic cell line.

Images